Fig. 3

Relationships between the change in rat sightings during the stay-at-home order in Chicago and proximity to restaurants (a) or housing type (b) based on ordinal regression of survey responses. Lines show the probability of a survey respondent reporting more rats (red solid line), fewer rats (blue dashed line), about the same number of rats (black dotted line) or no rats (gray dotted line) relative to a month prior. The count of restaurants was log-transformed for analysis due to right skewness and back-transformed for ease of interpretation. Shaded bands (a) or error bars (b) show 95% confidence intervals